How to eliminate texts from the data file?

I have a .dat file which has three columns of numbers (X,Y and Z). An example file is attached here. In the file there are some texts which prevents me from calling to matlab. In the beginning of the data, it is written "Measurement contains X,Y,Z)" and the next line "Injection #1"... after some rows of values it again says "Injection #2" .. again after some rows "Injection #3" so and so.... I want to eliminate these texts and call the whose rows in each column as X, Y and Z. Can anyone please tell me how to call the file and eliminate those texts in it?

PS: The attached one is just an example. In the real files, I have more than thousands of rows. So, impossible to delete manually.

[Cedric]

Does the file contain the first line "This measurement contains x,y and time" or is it something that you added? Do you need to read all data entries together or do you need to keep track of the injection number?

[aneps]

The file contains the text "This measurement contains x,y and time". An example file is attached with the question. Thank you.

Answer 1

[Cedric]

If the first line "This measurement contains x,y and time" is not present (if you added it for us to understand), you can do something as simple as:

 buffer = fileread('Test.txt') ;
 buffer = regexprep( buffer, 'i.*?\n', '\n') ;
 buffer = textscan(buffer, '%f %f %f') ;
 data   = [buffer{:}] ;

Let me know if you have to eliminate the first line. We can replace FILEREAD with a few lines involving FOPEN/FGETL/FREAD, which skip the first line and read the rest of the file.

EDIT after your comment.

 % - Read file, skip 1st line.
 fid = fopen('Test.txt', 'r') ;
 fgetl(fid) ;                                  % Skip 1st line.
 buffer = fread(fid, [1,Inf], '*char') ;
 fclose(fid) ;

 % - Eliminate 'injection' headers.
 buffer = regexprep( buffer, 'i.*?\n', '\n') ;

 % - Convert to numeric type.
 data = textscan(buffer, '%f %f %f') ;

Then you can extract columns they way you need them.. you can build vectors x, y, t as follows:

 x = data{1} ;
 y = data{2} ;
 t = data{3} ;

or, if you prefer, build a 3 columns array as follows

 data = [data{:}] ;

[aneps]

The file contains the text "This measurement contains x,y and time". I need to eliminate that as well. An example file is attached with the question. Thank you.

[Cedric]

See my edit.

[aneps]

Thanks a lot.

[aneps]

Hello, I am stuck with a few more things related to the same question. May I ask your help please? In the above question, how can I get the number of events (each x,y and time is an event, means each row is an event) after each 'injections'?

[Cedric]

Hi Aneps, if you still have this question, please send me an email, or post a new question and send me a link by email so I see it.

Well, what about the following actually? Look at data and counts, they contain all injections data by block and counts of events.

 content = fileread( 'Test.txt' ) ;
 blocks  = regexp( content, '#:\d+\s*([^i]+)', 'tokens' ) ;
 data    = cellfun( @(c) reshape(sscanf(c{1}, '%f'), 3, []).', blocks, ...
                    'UniformOutput', false ) ;
 counts  = cellfun( @(d) size(d, 1), data ) ;

[aneps]

Thanks a lot. I did program as you suggested, but it is not giving me what I am looking for. The first program you gave me is working wonderful, it reads the data the way I want and I made a long program from there

 .
 .
 .
 data = [data{:}] ;

I would like to add a few more lines to get more information. I have posted it as a new question (<http://www.mathworks.com/matlabcentral/answers/128448-counting-the-number-of-events-and-total-number-of-blocks-from-a-file)>. Thanks a lot.